In a way, what the article is describing is sort of like building a scientific hypothesis: your feedback needs to be concrete enough to be falsifiable.
You don't have to know all the facts, and you don't have to come across as overconfident. You should present the facts as you understand them. This gives the other person the ability to contextualise the feedback you're trying to give them, but also the necessary context to contradict you if they think the feedback is wrong. Likewise, putting it in terms of business outcomes means that they can counter your argument with what their intended outcome was, where appropriate (which is an especially important conversation with senior/high skilled people).
All the caveats you're describing are pretty damn important, but they're all pretty much part of what the author means by "be kind".
You don't have to know all the facts, and you don't have to come across as overconfident. You should present the facts as you understand them. This gives the other person the ability to contextualise the feedback you're trying to give them, but also the necessary context to contradict you if they think the feedback is wrong. Likewise, putting it in terms of business outcomes means that they can counter your argument with what their intended outcome was, where appropriate (which is an especially important conversation with senior/high skilled people).
All the caveats you're describing are pretty damn important, but they're all pretty much part of what the author means by "be kind".